Developmental pattern of calmodulin-binding proteins in rat jejunal epithelial cells

scapGNN: A graph neural network-based framework for active pathway and gene module inference from single-cell multi-omics data

Although advances in single-cell technologies have enabled the characterization of multiple omics profiles in individual cells, extracting functional and mechanistic insights from such information remains a major challenge. Here, we present scapGNN, a graph neural network (GNN)-based framework that creatively transforms sparse single-cell profile data into the stable gene-cell association network for inferring single-cell pathway activity scores and identifying cell phenotype-associated gene modules from single-cell multi-omics data. Systematic benchmarking demonstrated that scapGNN was more accurate, robust, and scalable than state-of-the-art methods in various downstream single-cell analyses such as cell denoising, batch effect removal, cell clustering, cell trajectory inference, and pathway or gene module identification. scapGNN was developed as a systematic R package that can be flexibly extended and enhanced for existing analysis processes. It provides a new analytical platform for studying single cells at the pathway and network levels.

Regulation and role of brush border-associated ERK1/2 in intestinal epithelial cells

We have recently shown that elevated extracellular signal-regulated kinase (ERK) activities stimulate proliferation of intestinal cells whereas low sustained levels of ERK activities correlate with Gl arrest and are required for expression of several enterocyte differentiation proteins. In an attempt to clarify how ERK1/2 regulates intestinal differentiation, the present study assessed the subcellular distribution and regulation of ERK proteins and activities in differentiated enterocytes. We report that (1) ERK1/2 and their upstream modulators Ras, p85 (PI-3K), Rac1, and MEK1 are found in the brush border; (2) brush border-associated ERK1/2 are stimulated by EGF and feeding; (3) immunoblotting of proteins phosphorylated on SP/K motif suggests the presence of ERK substrates in the brush border, one of which could be actin; and (4) pharmacological inhibition of ERK alters microvilli architecture. Our results suggest that ERK may play important roles in the control of microvilli structure and possibly, in brush border-associated responses in differentiated intestinal epithelial cells.

An in vitro model for the analysis of intestinal brush border assembly. II. Changes in expression and localization of brush border proteins during cell contact-induced brush border assembly in Caco-2BBe cells

In the companion paper (M. D. Peterson and M. S. Mooseker (1993). J. Cell Sci. 105, 445–460) we describe a method for modeling brush border assembly in the Caco-2BBe clones. In this study we have examined the molecular changes accompanying cell contact-induced brush border assembly. A subset of brush border proteins was tracked throughout brush border assembly by immunoblotting and by immunofluorescent localization using laser scanning confocal microscopy. Actin, fodrin, villin and presumptive unconventional myosin immunogens were distributed at the periphery of depolarized cells. All proteins partitioned primarily with the membrane fraction upon differential sedimentation of depolarized cell lysates; the fractionation patterns were comparable to those of confluent cells. After a monolayer had formed, each protein showed a redistribution to the apical domain in a discrete sequence. Actin and villin began to shift apically at 2 d, while fodrin and the unconventional myosin immunogens did not redistribute until 3 d. Enterocyte-like localization was observed by 5 d for all proteins. Sucrase-isomaltase was not reliably detectable until 9 d by immunofluorescence, after brush border assembly was complete. Quantitative immunoblot analysis of total cell extracts demonstrated an average 10-fold increase in villin levels, while fodrin levels appeared to remain unchanged. Three putative unconventional myosin immunogens of 140 kDa, 130 kDa, and 110 kDa have been detected previously in the C2BBe cells with a head-specific monoclonal antibody to avian brush border myosin I (M. D. Peterson and M. S. Mooseker (1992) J. Cell Sci. 102, 581–600). Each of these immunogens displayed distinct expression patterns during brush border assembly. The 140 kDa species decreased by half, while the 130 kDa immunogen(s) did not change in any consistent fashion. The 110 kDa protein, presumed to be human brush border myosin I, rose on average 8-fold. A ribonuclease protection assay was also performed using a probe for human brush border myosin I. Equal amounts of total RNA from depolarized and confluent cells were assayed; the level of protected product was approximately 9-fold greater in the confluent cells. The expression patterns of the brush border proteins, coupled with the correlation to the ultrastructural features during brush border assembly in C2BBe cells, show that differentiation of the C2BBe cells closely resembles the changes that occur during human fetal intestinal differentiation.

Identification, characterization and cloning of myr 1, a mammalian myosin-I

We have identified, characterized and cloned a novel mammalian myosin-I motor-molecule, called myr 1 (myosin-I from rat). Myr 1 exists in three alternative splice forms: myr 1a, myr 1b, and myr 1c. These splice forms differ in their numbers of putative calmodulin/light chain binding sites. Myr 1a-c were selectively released by ATP, bound in a nucleotide-dependent manner to F-actin and exhibited amino acid sequences characteristic of myosin-I motor domains. In addition to the motor domain, they contained a regulatory domain with up to six putative calmodulin/light chain binding sites and a tail domain. The tail domain exhibited 47% amino acid sequence identity to the brush border myosin-I tail domain, demonstrating that myr 1 is related to the only other mammalian myosin-I motor molecule that has been characterized so far. In contrast to brush border myosin-I which is expressed in mature enterocytes, myr 1 splice forms were differentially expressed in all tested tissues. Therefore, myr 1 is the first mammalian myosin-I motor molecule with a widespread tissue distribution in neonatal and adult tissues. The myr 1a splice form was preferentially expressed in neuronal tissues. Its expression was developmentally regulated during rat forebrain ontogeny and subcellular fractionation revealed an enrichment in purified growth cone particles, data consistent with a role for myr 1a in neuronal development.

Changes in keratin expression during fetal and postnatal development of intestinal epithelial cells

We have investigated keratin expression in fetal, newborn and adult rat intestines by immunofluorescence staining, immunoblotting of two-dimensional gels and Northern blot analysis of total cellular RNAs. Keratin-type intermediate filaments, composed predominantly of keratin no. 19, were observed already in the undifferentiated stratified epithelium present at 15-16 days of gestation. The marked maturation and differentiation of the epithelium taking place at 18-19 days of gestation was characterized by the appearance of the differentiation-specific keratin no. 21 and by a significant increase in the relative amount of keratin no. 8. The keratin pattern typical of adult villus cells became established at the time of birth, and was marked by a considerable increase in the complexity of the keratin-related polypeptides detected on two-dimensional gels, indicative of extensive post-translational modification of all keratins. Starting at 20 days of gestation there was a major increase in the relative abundance of mRNAs coding for keratin nos. 8, 19 and 21; in contrast, the relative amount of keratin no. 18 mRNA reached a peak shortly after birth and declined to very low levels in adult intestine. These results demonstrated marked changes in keratin expression and post-translational processing taking place at key stages of intestinal development. The appearance of keratin no. 21 in coincidence with the formation of an adult-type brush border and terminal web would be consistent with it having an important role in the organization of the intermediate filament network in the apical cytoplasm of the differentiated intestinal cells.

Limited tissue distribution of the intestinal brush border myosin I protein

A myosinlike 105-110-kilodalton calmodulin-binding protein, brush border myosin I, found in the intestinal brush border has been linked to two seemingly disparate but possibly interacting functions of the brush border, namely, microvillar motility and vitamin D regulated calcium transport. If brush border myosin I were to function primarily as a myosinlike molecule powering cellular or microvillar motility, one might expect it to be found in a variety of tissues with microvilli such as the renal brush border and bile canaliculus. On the other hand, a more specialized function such as participation in vitamin D regulated calcium transport might dictate a more restricted tissue distribution for brush border myosin I. To determine the tissue distribution of brush border myosin I, we purified this protein to apparent homogeneity, generated antisera to it, and used the antisera to localize the protein within the intestinal epithelial cell by immunocytochemistry. We then screened a variety of other tissues (brain, lung, heart, liver, spleen, pancreas, kidney, and skeletal muscle) both for calmodulin-binding proteins as well as for brush border myosin I using Western blots and immunofluorescence. Our results indicate that the intestinal brush border myosin I is limited in its distribution to the intestinal brush border.

Assembly of the brush border cytoskeleton: changes in the distribution of microvillar core proteins during enterocyte differentiation in adult chicken intestine

The assembly of the intestinal microvillus cytoskeleton was examined during the differentiation of enterocytes along the crypt-villus axis in adult chicken duodenum using light and electron microscopic immunolocalization techniques. Using antibodies reactive with villin, fimbrin, and the heavy chain (hc) of brush border (BB) myosin I (110K-calmodulin complex) and rhodamine-conjugated phalloidin as a probe for F-actin, we determined that while actin, villin, and fimbrin were all localized apically along the entire axis, BB myosin I (hc) did not assume this localization until the crypt-villus transition zone. In addition to their localization at the BB surface, all four proteins were present at significant levels along the lateral margins of enterocytes along the entire crypt-villus axis, suggesting that these proteins may be involved in the organization and function of the basolateral membrane cytoskeleton as well. The pattern of expression of the microvillar core proteins along the crypt-villus axis in the adult was comparable to that seen in the intestine of the late stage chicken embryo and suggests that a common program for brush border assembly may be used in both modes of enterocyte differentiation.

Structural and immunological characterization of the myosin-like 110-kD subunit of the intestinal microvillar 110K-calmodulin complex: evidence for discrete myosin head and calmodulin-binding domains

The actin bundle within each microvillus of the intestinal brush border is tethered laterally to the membrane by spirally arranged bridges. These bridges are thought to be composed of a protein complex consisting of a 110-kD subunit and multiple molecules of bound calmodulin (CM). Recent studies indicate that this complex, termed 110K-CM, is myosin-like with respect to its actin binding and ATPase properties. In this study, possible structural similarity between the 110-kD subunit and myosin was examined using two sets of mAbs; one was generated against Acanthamoeba myosin II and the other against the 110-kD subunit of avian 110K-CM. The myosin II mAbs had been shown previously to be cross-reactive with skeletal muscle myosin, with the epitope(s) localized to the 50-kD tryptic fragment of the subfragment-1 (S1) domain. The 110K mAbs (CX 1-5) reacted with the 110-kD subunit as well as with the heavy chain of skeletal but not with that of smooth or brush border myosin. All five of these 110K mAbs reacted with the 25-kD, NH2-terminal tryptic fragment of chicken skeletal S1, which contains the ATP-binding site of myosin. Similar tryptic digestion of 110K-CM revealed that these five mAbs all reacted with a 36-kD fragment of 110K (as well as larger 90- and 54-kD fragments) which by photoaffinity labeling was shown to contain the ATP-binding site(s) of the 110K subunit. CM binding to these same tryptic digests of 110K-CM revealed that only the 90-kD fragment retained both ATP- and CM-binding domains. CM binding was observed to several tryptic fragments of 60, 40, 29, and 18 kD, none of which contain the myosin head epitopes. These results suggest structural similarity between the 110K and myosin S1, including those domains involved in ATP- and actin binding, and provide additional evidence that 110K-CM is a myosin. These studies also support the results of Coluccio and Bretscher (1988. J. Cell Biol. 106:367-373) that the calmodulin-binding site(s) and the myosin head region of the 110-kD subunit lie in discrete functional domains of the molecule.

Expression of brush border calmodulin-binding proteins during human small and large bowel differentiation

The expression and immunocytochemical localization of three brush border cytoskeletal calmodulin-binding proteins, caldesmon, fodrin, and the 110 kDa subunit of the 110 kDa calmodulin complex, have been studied in human intestinal epithelial cells as a function of their ontogenic differentiation. At immature stages (fetal week 8), caldesmon and fodrin were present in undifferentiated intestinal epithelial cells. However, no 110 kDa protein was detectable except a 135 kDa immunoreactive species. The 110 kDa form appeared at week 12, when microvilli differentiate, and became prominent at week 14 simultaneously with the disappearance of the 135 kDa species. Finally at week 14, the calmodulin-binding protein pattern was identical to that found in adults. Immunocytochemical experiments revealed that at week 8, antibodies to caldesmon and fodrin gave a fluorescence lining at the periphery of the cells, whereas the 110 kDa immunoreactive species was hardly detectable. Then, as early as week 12 of gestation, with the three antisera, a bright fluorescence lined the apex of the cells, as in adults. In the colon, the events were delayed. This study demonstrates that the developmental pattern of the three calmodulin-binding proteins investigated, caldesmon, fodrin and the 110 kDa subunit, parallels the temporal differentiation of human intestinal brush borders and the proximo-distal morphological intestinal maturation.

Calmodulin in normal and cystic fibrosis human intestine at different developmental stages

Calmodulin concentrations and localisation have been analysed as a function of development in human intestinal epithelial cells from normal and cystic fibrosis individuals. In normal fetuses up to eight weeks of gestation intestinal epithelial cells which were still undifferentiated were not immunoreactive and their calmodulin content was low. From eight weeks onwards there was a significant overall increase in calmodulin content concomitant with its segregation to the apical side of epithelial cells. At 14 weeks of gestation calmodulin concentrations and localisation closely resembled those of adults. The developmental pattern of calmodulin appeared to parallel the morphological and functional maturation of brush borders which occurs during the first trimester of pregnancy. In the intestinal epithelial cells from a 19 weeks cystic fibrosis fetus and a cystic fibrosis newborn infant neither calmodulin concentration, nor its localisation were affected. Similarly, brush border calmodulin binding proteins and enzymatic activities were similar in normal subjects and the cystic fibrosis intestine.